In many agile setups inspired by Spotify-style squads, you often see a pattern where squads rely on one architect as the go-to person for every significant decision, which turns that architect into a bottleneck. In that kind of setup, pretty much every important architecture decision and meaningful change ends up back on that one architect’s desk. That arrangement is manageable while the programme is small, because the volume and complexity of decisions stay within what one architect can realistically handle. But as the programme expands, relying on a single architect for the big decisions becomes unsustainable, and the way of working starts to strain and falter. What happens next is exactly the point of this piece. Instead of trying to make every decision, the architect focuses on creating a repeatable way for squads to make solid decisions on their own.
That repeatable way of working has a name. Most practices call it an architecture playbook, yet many of them also have one that people rarely use.
An architecture playbook is the operating manual of an architecture practice. It’s designed to carry your agreed way of working from one engagement to the next, so your good practices don’t disappear when you move on from the project or the organisation. It sounds simple in theory, yet many capable architecture teams struggle to create a playbook that sticks, and others find that their carefully written playbooks are rarely used day-to-day. So why is it that even experienced architects find this hard, and why do architecture playbooks so often end up underused or ignored? I went back to the research to find out.
What an architecture playbook actually is
Let’s start with what a playbook isn’t, because the wrong mental picture is often why architects overlook its value. Writing a playbook and then keeping it up to date can easily look like routine work rather than strategic architecture activity. In many teams, this kind of documentation is treated as unglamorous admin work that follows delivery, rather than as part of the core architecture effort. So it quietly drops down the priority list, squeezed out by deadlines and delivery pressure, until it barely shows up on the list at all.
That instinct makes sense in fast-paced delivery environments, yet it’s precisely what leads to a misunderstanding of what a playbook can actually do for the practice.
A playbook isn’t just a stack of architecture diagrams and design blueprints. It’s not simply a wiki full of standards. It’s not a governance folder you complete after the fact just to show that a decision was made. All of those artefacts exist and have their place, but they’re not what I mean when I talk about a playbook.
A strong playbook carries the principles that shape design decisions, the end-to-end workflows for structuring work, the governance that keeps delivery aligned with the architecture, and the implementation practices that translate architectural intent into working solutions. It serves as a how-to layer that turns static architectural designs into repeatable working practices that teams can adopt.
And that sets the stage for an important question we need to address clearly. Should we think of a playbook as just a document, or as part of an ongoing process? It’s actually both, and that’s where many teams miss the mark. Ahlemann and colleagues, studying where enterprise architecture management really pays off, argue from a resource-based perspective that artefacts alone create limited value unless the organisation builds capabilities to model, plan, implement, and govern using them. A playbook that you write once and then shelve isn’t just incomplete; it’s effectively useless. Writing the playbook is the easy part. Building the organisational capability around it is the more demanding part. It’s the practice around the playbook, the people who maintain it and actually use it, that makes it worth anything.
I am living this at the moment. I’m helping set up an architecture playbook working group, and we meet every week to work through standards, processes, and reference architectures. It’s slow, it’s not glamorous, and yet it’s some of the most useful work we do. That’s because the playbook isn’t just the pages we write in those meetings. It is the shared way of working we are slowly building, the kind that outlives any one of us. Architects eventually move on. Team structures and membership also evolve. If the way of working only lives in people’s heads, it walks out the door when they do. And the whole point of a playbook is to make sure that doesn’t happen.
Why the one you already built goes unread
In a lot of organisations, enterprise architecture is approached more as a documentation task than as a sustained way of working embedded in the operating model. They’re keen to get the artefacts, the diagrams, blueprints, and models, but they don’t invest in the practice needed to keep those artefacts alive and useful. In some organisations, enterprise architecture exists mostly as a label, with limited dedicated staffing for the discipline itself. In many cases, organisations assign the ‘enterprise architect’ title to solution or delivery architects whose responsibilities remain focused on individual projects rather than enterprise-level concerns. In that setup, most architectural roles are incentivised around near-term delivery, while playbook development is a longer-term investment that is not in their scope. And the payoff, greater agility and a practice that can scale, belongs to the practice as a whole rather than to any single project, so nobody in a project seat has an obvious reason to build it.
So the artefacts still get produced, because someone asks for them, and then they sit there. The very people who are supposed to use those artefacts don’t reach for them. This isn’t about laziness. It’s that under tight deadlines, delivery architects fall back on familiar methods. When they haven’t been involved in shaping a playbook, they neither trust it nor have time to interpret it, especially if it seems unnecessary to complete the immediate project. The playbook is answering questions about the practice, not about their project. How do you scale the architecture work itself? How do you produce models faster, in a repeatable way?
Underneath all of this are a few specific failure patterns, and I’ve seen every one of them play out.
First, the framework itself is often too heavy. It’s theoretical, hard to follow, and sometimes impossible to apply as written. When a document is that dense, a busy architect won’t reach for it under deadline. So they tend to rely on personal experience and memory rather than the playbook. The version people really use is short enough to keep in your head and specific enough to answer a real question.
Another common issue is that playbooks are usually developed as separate initiatives, rather than being embedded into actual planning and delivery activities. It turns into a separate architecture programme that runs alongside planning and delivery, the places where decisions really get made. So it never enters the room where those decisions happen. Since playbooks are most relevant at decision points, they need to be embedded in the tools and processes used in those moments, rather than stored in rarely accessed repositories.
In many cases, nobody has agreed who should use the playbook, when, or for what purpose. There’s no governance around the playbook itself. Niemi and Pekkola, in their research on EA artefacts, highlight that organisations often fail to define why, how, when, and by whom those artefacts should be used. Nobody clearly defines the process, so they just sit there without a moment they really belong to. Their case study identified multiple distinct situations in which EA artefacts get used, so those decision moments do exist. They are simply never named.
And then there’s sponsorship, which is often thin. This is the one I have seen sink more practices than any framework choice. Without a senior sponsor who expects the playbook to be used and asks whether it was, following it remains optional. People default to short-term issues and deadlines, which is ironic, because the playbook is exactly what would make those deadlines easier to hit. It is your operating manual, your repeatable way of working. Skip it to save time this week, and you’ll end up rebuilding from scratch next week, and the week after that. Teams duplicate work that already exists, run into entirely predictable integration problems, and keep re-solving the same issues because the original reasoning was never written down anywhere the next team could find it.
Put plainly, many unread playbooks aren’t unread because they’re badly written. They are unread because they are badly owned, badly placed, and built to be filed rather than followed.
What to do about it
On multiple engagements, I found myself repeatedly applying the same approach. Focus on business capabilities first, and only then consider the supporting technology. Project after project would start leaning toward a particular product before anyone had really understood the business capability it was meant to serve. My role repeatedly involved redirecting the conversation to the underlying business context. Pause to identify the business capability clearly, and only afterwards assess which products (applications) could best support it.
After applying this approach repeatedly, it began to feel less like situational judgement and more like a consistent principle. And that illustrates the point: When an instinctive approach repeatedly leads to good outcomes, it represents reasoning that a playbook is designed to capture and share. Documented as a principle, this approach becomes reusable guidance: identify the business capability first and allow it to drive subsequent technology choices, rather than starting from products. Put that principle in the playbook, with the reasoning beside it, and the next architect doesn’t have to rediscover it under pressure on their own engagement. This is just one example, and every organisation has its own way of doing architecture, shaped by its context and culture, with distinctive nuances. The playbook exists to capture and communicate those organisation-specific ways of working.
In many organisations, reasoning like that never gets written down. It gets developed during high-pressure engagements and then lost when the architect or the team moves on, because nobody wrote it down. The remedy is to treat the playbook as a practical tool for daily use, rather than as a static deliverable, and several practical actions help turn that intention into reality.
First, integrate the playbook into existing planning and delivery activities rather than treating it as a separate stream of work. It must be accessible within the tools and forums where planning and decisions happen. Otherwise it won’t be consulted at the critical moment.
Second, keep it lean, on purpose. Documenting every detail exhaustively burns time and money, and the more detailed the artefact, the sooner it goes out of date. Make it easy for the user. A concise playbook people reach for beats a comprehensive one that sits unused.
Third, capture the reasoning while it’s still fresh. Write down the ‘why’, the constraints you were under, and the options you turned down, and do it there and then, because that’s the only moment it’s really worth anything. Come back six months later and you’re reconstructing context you’ve already lost. You’re building a repeatable way of working, so make it reliable by writing the reasoning down promptly.
Fourth, be explicit about who is responsible for maintaining the playbook, who should use it, and who is backing it from the top. This is where organisational culture turns into concrete practice. The reasons a playbook goes unread are mostly cultural, a matter of priorities, incentives, and habits, and none of those change on their own. Culture shifts when someone senior expects the playbook to be used and asks whether it was, making adherence the natural, low-friction choice instead of an optional extra.
A playbook doesn’t scale a practice on its own. Effective scaling relies on your engagement model, the governance mechanisms you put in place, and the people you build and their competencies. What the playbook does is make every engagement cheaper to run, through consistency and faster onboarding so that you can take on more without the quality dropping. That is a tangible contribution to scale, but it’s not the whole story. If you invest heavily in documentation while neglecting people, governance, and operating models, it undermines the very purpose of having a playbook.
The measure that matters
Remember those squads routing every decision through one architect? Practices that get out of that bottleneck do not do it by adding more documentation. They do it by changing how decisions get made. The architect stops being the person every important decision runs through, and the way of working moves out of one head and into something the squads can pick up and use themselves. That is the fix. Not more documentation, but a practice that no longer needs a particular person in the room, and that spreads architectural thinking across the teams, including the people who are not architects. They build a playbook.
The test for your own playbook is whether it enables good work without your direct involvement. Not how complete it looks, not how many pages it runs to. If every real decision still waits for you, you don’t really have a playbook. You have a bottleneck with good documentation.
I write The Foresight-Driven Enterprise for architects who are building practices meant to outlast them. If this resonated with you, check out other articles in my newsletter.
Further Reading
Ahlemann, Legner and Lux (2021), A resource-based perspective of value generation through enterprise architecture management, Information and Management.
Kotusev and Kurnia (2020), The theoretical basis of enterprise architecture: A critical review and taxonomy of relevant theories, Journal of Information Technology.
Löhe and Legner (2013), Overcoming implementation challenges in enterprise architecture management: a design theory for architecture-driven IT Management (ADRIMA), Information Systems and e-Business Management.
Niemi and Pekkola (2017), Using enterprise architecture artefacts in an organisation, Enterprise Information Systems.
Embodied practice is living practice. The book lives in the practice, as it is played. Artifacts as memory aids, shadows of the practice, not as the thing itself. When the artifact becomes “the playbook”, the play gets frozen and forgotten. I am tending to think the playbook exists in the process of keeping it alive and current, it is just an anchor point for humans to meet and sync their mental models. So, a live, interacting, Community of Practice, is the book that plays in the field.
Many thanks for this insightful piece.